Calcium signalling silencing in atrial fibrillation

Abstract

Subcellular calcium signalling silencing is a novel and distinct cellular and molecular adaptive response to rapid cardiac activation. Calcium signalling silencing develops during short‐term sustained rapid atrial activation as seen clinically during paroxysmal atrial fibrillation (AF). It is the first ‘anti‐arrhythmic’ adaptive response in the setting of AF and appears to counteract the maladaptive changes that lead to intracellular Ca²⁺ signalling instability and Ca²⁺‐based arrhythmogenicity. Calcium signalling silencing results in a failed propagation of the [Ca²⁺]_i signal to the myocyte centre both in patients with AF and in a rabbit model. This adaptive mechanism leads to a substantial reduction in the expression levels of calcium release channels (ryanodine receptors, RyR2) in the sarcoplasmic reticulum, and the frequency of Ca²⁺ sparks and arrhythmogenic Ca²⁺ waves remains low. Less Ca²⁺ release per [Ca²⁺]_i transient, increased fast Ca²⁺ buffering strength, shortened action potentials and reduced L‐type Ca²⁺ current contribute to a substantial reduction of intracellular [Na⁺]. These features of Ca²⁺ signalling silencing are distinct and in contrast to the changes attributed to Ca²⁺‐based arrhythmogenicity. Some features of Ca²⁺ signalling silencing prevail in human AF suggesting that the Ca²⁺ signalling ‘phenotype’ in AF is a sum of Ca²⁺ stabilizing (Ca²⁺ signalling silencing) and Ca²⁺ destabilizing (arrhythmogenic unstable Ca²⁺ signalling) factors. Calcium signalling silencing is a part of the mechanisms that contribute to the natural progression of AF and may limit the role of Ca²⁺‐based arrhythmogenicity after the onset of AF.

Abbreviations

AF

atrial fibrillation

[Ca²⁺]_i

intracellular calcium concentration

CaMKII

Ca²⁺/calmodulin dependant kinase II

CaT

calcium transient

CRU

calcium release unit

DAD

delayed afterdepolarization

EAD

early afterdepolarization

LTTC

voltage gated L‐type Ca²⁺ channel

I_Ca,L

L‐type Ca²⁺ current

IP₃R

inositol‐1,4,5‐trisphosphate‐receptor

I_TI

transient inward current

[Na⁺]_i

intracellular sodium concentration

NCX

Na⁺/Ca²⁺ exchanger

PKA

protein Kinase A

RyR2

ryanodine receptor type 2

SR

sarcoplasmic reticulum

TT

T‐tubule

Introduction

Atrial fibrillation (AF) is the most common arrhythmia, with a prevalence of > 6% in the population (Kannel et al. 1998). Despite advances in understanding the underlying disease mechanisms and the development of more effective treatment options, prevalence of AF is projected to increase even further (Kirchhof et al. 2013). Recent guidelines advocate a ‘personalized’ management of AF, which focuses on treating AF according to individual disease mechanisms (Kirchhof et al. 2013). It is becoming more evident that the different aetiologies of AF lead to distinct ‘types’ of AF. This implies that atrial fibrillation is likely a common endpoint for disparate pathomechanisms. It is therefore crucial to better understand the mechanisms underlying these different types of AF to be able to offer treatment and/or preventative measures specifically tailored to a type of AF.

Subcellular Ca²⁺ signalling undergoes complex changes as part of the atrial remodelling process before and/or after the initiation of the arrhythmia (Greiser et al. 2011). It is still contentious, however, which role these changes play in AF arrhythmogenesis. Specifically, the contribution of Ca²⁺‐based arrhythmogenic mechanisms to AF initiation and maintenance remains controversial. The characterization of the novel mechanism of subcellular Ca²⁺ signalling silencing shows that the rapid atrial rate, which occurs after the onset of AF, initiates an adaptive protective response rather than arrhythmogenic Ca²⁺ instability. Ca²⁺ signalling silencing counteracts the rapid rate‐induced intracellular Ca²⁺ overload and results in an anti‐arrhythmic stabilization of subcellular Ca²⁺ signalling.

There is evidence that unstable intracellular Ca²⁺ signalling occurs in AF (Hove‐Madsen et al. 2004; Neef et al. 2010). This Ca²⁺ signalling instability has been suggested to contribute to atrial arrhythmogenicity and the maintenance of AF in patients (Voigt et al. 2012). The discovery of Ca²⁺ signalling silencing suggests that the Ca²⁺ signalling ‘phenotype’ present in a specific AF patient is the sum of Ca²⁺ stabilizing (i.e. Ca²⁺ signalling silencing) and Ca²⁺ destabilizing (i.e. arrhythmogenic Ca²⁺ instability) factors. Consistent with this concept is the finding that in some patients the effects of Ca²⁺ signalling silencing appear to prevail (Greiser et al. 2014). The identification of the Ca²⁺ silencing response has important implications for clinical AF and the role of Ca²⁺‐based arrhythmogenesis in the natural progression of AF.

Normal Ca²⁺ handling in atrial myocytes

Membrane depolarization activates voltage‐gated L‐type Ca²⁺ channels (LTCCs). This initiates the influx of a small amount of Ca²⁺ into the cardiac myocyte. The small increase in subsarcolemmal Ca²⁺ triggers the release of larger quantities of Ca²⁺ from the nearby Ca²⁺ storage organelle of the cardiac myocyte, the sarcoplasmic reticulum (SR), through its Ca²⁺ release channels, the ryanodine receptors (RyR2). This transient increase in intracellular calcium concentration ([Ca²⁺]_i) initiates contraction as free Ca²⁺ binds to the myofilaments. This Ca²⁺‐induced Ca²⁺ release, which is characteristic for cardiac excitation–contraction coupling, is fundamentally the same in atrial and ventricular myocytes (Bers, 2001). Relaxation occurs as Ca²⁺ is released from the myofilaments, re‐sequestered into the SR by the SR Ca²⁺‐ATPase and extruded from the cell by the sarcolemmal Na⁺/Ca²⁺ exchanger (NCX).

Atrial calcium release

Calcium release units (CRUs) consist of groups of RyR2s that are formed in specialized regions of the SR (Franzini‐Armstrong et al. 1999). Upon activation this primary Ca²⁺ release unit produces a Ca²⁺ spark, which is the fundamental Ca²⁺ release event (Cheng et al. 1993). In atrial myocytes CRUs are organized along the Z‐disk of the sarcomere and spaced ∼1 μm apart, similar to CRU organization in ventricular myocytes (Chen‐Izu et al. 2006).

Important modulation of the RyR2s’ sensitivity to triggering by [Ca²⁺]_i is achieved through phosphorylation, nitrosylation and oxidation (O'Brien et al. 2015). Higher RyR2 Ca²⁺ sensitivity has been implicated in contributing to unstable arrhythmogenic Ca²⁺ signalling in ventricular (heart failure) and atrial (atrial fibrillation) disease (see below) and a distinct alteration in RyR2 phosphorylation is part of Ca²⁺ signalling silencing. While RyR2s are the major Ca²⁺ release channels in cardiac myocytes, Ca²⁺ release in atrial myocytes appears to be modulated by inositol‐1,4,5‐trisphosphate‐receptors (IP₃Rs), which colocalize to RyR2s and are much more abundant in atrial than ventricular myocytes (Lipp et al. 2000). IP₃Rs are activated by IP₃ that is generated in response to G protein‐coupled receptor or receptor tyrosine kinase stimulation. The precise effect of IP₃R‐mediated modulation of atrial Ca²⁺ release is controversial in the literature. However, knockout of the predominant cardiac isoform of IP₃R (type 2) abolished an endothelin 1‐mediated positive inotropic and arrhythmogenic effect (Li et al. 2005), suggesting that IP₃Rs can be potent modulators of atrial calcium signalling.

Cellular arrhythmogenic mechanisms: Ca²⁺ waves and ectopic focal discharges

Ca²⁺‐based arrhythmogenicity caused by unstable intracellular Ca²⁺ signalling is mediated by (1) intracellular Ca²⁺ waves and (2) Ca²⁺‐activated inward current. Ca²⁺ ‘overload’ (i.e. increased [Ca²⁺]_SR) (Wier et al. 1987) increases the sensitivity of the RyR2s to be activated by cytosolic [Ca²⁺]_i. This higher [Ca²⁺]_i sensitivity leads to an increased probability of spontaneous Ca²⁺ sparks and Ca²⁺ waves. Under these conditions, in addition to the Ca²⁺ release initiated by an action potential (AP), there can be Ca²⁺ release events during diastole between the normal AP initiated events. When [Ca²⁺]_i increases due to spontaneous Ca²⁺ release events, a Ca²⁺‐based membrane current is activated during diastole. This arrhythmogenic transient inward current (I _TI), which is carried by the NCX, is responsible for the generation of delayed afterdepolarizations (DADs) (Berlin et al. 1989). When DADs are sufficiently large they can trigger an extrasystole. Additionally, Ca²⁺‐activated I _TI can occur during repolarization and then contribute to early afterdepolarizations (EADs), which can also trigger extrasystoles. Also, during short APs in the presence of a high [Ca²⁺]_SR, two conditions found during AF, NCX‐mediated late phase 3 EADs can occur (Patterson et al. 2007).

Calcium signalling silencing: an anti‐arrhythmic response to rapid activation

Calcium signalling silencing constitutes the atrial Ca²⁺ signalling responses to a sustained high atrial activation rate in the absence of cardiovascular disease. Surprisingly, sustained rapid atrial activation does not lead to arrhythmogenic intracellular Ca²⁺ signalling in healthy atria. Instead, rapid rate alone produces Ca²⁺ signalling silencing. This mechanism acts to mitigate the effects of the high atrial rate with respect to Ca²⁺ instability. Key features of Ca²⁺ signalling silencing are (1) the failure of the centripetal intracellular Ca²⁺ wave propagation, which results in significantly reduced central‐cellular Ca²⁺ release, which underlies a reduced whole cell Ca²⁺ transient; (2) distinct remodelling of the RyR2 complex including a substantial reduction of RyR2 protein expression and reduced CaMKII‐mediated RyR2 phosphorylation; and (3) low intracellular sodium concentration ([Na⁺]_i), which further contributes to the ‘unloading’ of Ca²⁺. The above mechanisms lead to an unchanged (compared to controls) low rate of Ca²⁺ sparks and arrhythmogenic Ca²⁺ waves, another important feature of Ca²⁺ signalling silencing. (For an overview of changes in key Ca²⁺ handling proteins in Ca²⁺ signalling silencing, see Fig. 1.) Taken together, Ca²⁺ signalling silencing prevents arrhythmogenic intracellular Ca²⁺ signalling instability via novel and distinct mechanisms.

Figure 1. Schematic depiction of changes in function, expression and phosphorylation levels of key Ca²⁺ handling proteins in Ca²⁺ signalling silencing.

Arrows denote direction of change; see text for details. CaM, calmodulin; CaMKII, Ca²⁺/calmodulin‐dependent kinase II; I _CaL, L‐type Ca²⁺ current; MyBP‐C, myosin‐binding protein C; NCX, Na⁺/Ca²⁺ exchanger; NKA, Na⁺/K⁺‐ATPase; P, phosphate group; PKA, protein kinase A; PLB, phospholamban; PLM, phospholemman; PMCA, plasma membrane Ca²⁺‐ATPase; RyR2, ryanodine receptor type 2; Ser, serine residue; Serca2a, SR Ca²⁺‐ATPase type 2a; SR, sarcoplasmic reticulum; Thr, threonine residue; TnI, troponin I; TnC, troponin C.

Failure of the intracellular Ca²⁺ wave

Increased fast Ca²⁺ buffering, NCX upregulation and reduced L‐type Ca²⁺ current

Intracellular fast Ca²⁺ buffering strength is increased in Ca²⁺ signalling silencing due to increases in calmodulin protein expression and increased myofilament Ca²⁺ sensitivity (reduced troponin I phosphorylation, Fig. 1) (Greiser et al. 2014). Atrial cells without T‐tubules rely on Ca²⁺ diffusion between neighbouring RyR2 clusters to activate regenerative Ca²⁺ release from the periphery to the core of the myocyte (Berlin, 1995; Hüser et al. 1996). The higher fast Ca²⁺ buffering strength impedes Ca²⁺ diffusion in Ca²⁺ signalling silencing and contributes to the failure of the intracellular Ca²⁺ wave. The increased sensitivity of NCX to [Ca²⁺]_i seems to be a beneficial mechanism when myocytes that have undergone Ca²⁺ signalling silencing are rapidly activated. With high rate, Ca²⁺ entry will be greater due to the high activation frequency of the L‐type Ca²⁺ current (I _Ca,L) even when I _Ca,L amplitude is reduced. With the increased Ca²⁺ entry, NCX is better able to balance that high influx with a relatively higher efflux of Ca²⁺ than would be the case if the [Ca²⁺]_i sensitivity of the NCX were the same as it is in normal cells. This higher efflux of Ca²⁺ is reflected in the shortened decay of the subsarcolemmal Ca²⁺ transient (CaT), which contributes to the failure of the intracellular Ca²⁺ wave by reducing the subsarcolemmal trigger Ca²⁺ for the centripetal intracellular Ca²⁺ wave. In Ca²⁺ signalling silencing the unique combination of increased Ca²⁺ buffering strength, upregulated NCX function, reduced I _Ca,L and unchanged Ca²⁺ release properties of the SR (see below) contribute to the failure of the intracellular Ca²⁺ wave propagation.

Remodelling of the ryanodine receptor complex and arrhythmogenic Ca²⁺ release

Calcium signalling silencing leads to a distinct remodelling of the RyR2 complex. This entails a substantial and unique reduction of RyR2 protein expression (by ∼80%), increased RyR2 phosphorylation at a predominant PKA site (Ser2809) and reduced phosphorylation at a predominant CaMKII site (Ser2814). Additionally, expression of CaMKIIδ and auto‐phosphorylated CaMKII (CaMKIIpTh286) remains at normal levels (Greiser et al. 2014). While a moderate (20–40%) reduction in RyR2 protein expression levels is reported in most studies in AF patients and AF models (Voigt et al. 2012), it is not as substantial as in calcium signalling silencing. This has important functional implications for Ca²⁺ spark behaviour. As the RyR2 cluster grid is unchanged in calcium signalling silencing (Greiser et al. 2014), it can be deduced that the individual RyR2 clusters have a significantly reduced number of RyR2 channels. Simulations using an atrial cell model show that an 80% reduction in the number of RyR2 channels in a cluster results in a similarly significant reduction in Ca²⁺ spark frequency (Greiser et al. 2014). Reduced Ca²⁺ spark frequency resulting from ‘depleted’ RyR2 clusters can be offset by higher levels of RyR2 phosphorylation, which increase RyR2 open probability (Carter et al. 2006) and will therefore increase Ca²⁺ spark frequency. Increased RyR2 phosphorylation is a characteristic finding in human AF (Vest et al. 2005; El‐Armouche et al. 2006; Voigt et al. 2012) and AF models (Greiser et al. 2009). Phosphorylation‐mediated RyR2 sensitization has been implicated in underlying unstable Ca²⁺ signalling (i.e. increased Ca²⁺ spark frequency and diastolic Ca²⁺ leak) in AF. Indeed, high CaMKII expression levels are consistently found in atria of AF patients (Tessier et al. 1999; Neef et al. 2010). Moreover, a CaMKII inhibitor normalized increased Ca²⁺ spark frequency and diastolic Ca²⁺ leak in human atrial myocytes from AF patients with increased RyR2 phosphorylation at a predominant CaMKII site (Ser2814) (Neef et al. 2010) and reduced the number of DADs recorded from atrial myocytes from patients with AF (Voigt et al. 2012). In calcium signalling silencing, however, the effect of high RyR2 phosphorylation levels at a predominant PKA site (Ser2809) on Ca²⁺ spark frequency was mitigated by the substantial RyR2 protein depletion resulting in unchanged low Ca²⁺ spark and Ca²⁺ wave rates ([Ca²⁺]_SR is unchanged) (Greiser et al. 2014). While increased RyR2 phosphorylation levels at Ser2809 have been found to increase RyR2 open probability (Carter et al. 2006), other reports indicate Ser2030 as the dominant RyR2 PKA phosphorylation site (Xiao et al. 2005). Thus, there is some controversy as to the role of PKA‐mediated RyR2 phosphorylation, but it appears that increased levels of RyR2 phosphorylation at the predominant CaMKII site at Ser2814 facilitate arrhythmogenic unstable Ca²⁺ signalling in atrial myocytes. Rapid rate, which had been implicated in the activation of CaMKII in AF (Neef et al. 2010), does not induce increases in CaMKII expression in healthy atria and, importantly, even reduced RyR2 phosphorylation levels at the crucial CaMKII site (Ser2814). Thus, the ‘non‐activation’ of CaMKII by rapid rate in healthy atria is an important component of the calcium signalling silencing mechanism. It further suggests that activation of CaMKII in AF and subsequent increases in RyR2 phosphorylation are likely due to other mechanisms, most likely pre‐existing cardiovascular disease (e.g. cardiomyopathy, heart failure).

Finally, IP₃R protein expression (Yamada et al. 2001) and agonists (endothelin‐1, angiotensin II) in the IP₃ signalling cascade are upregulated in AF (Goette et al. 2000; Mayyas et al. 2010). The role for IP₃R‐mediated Ca²⁺ release in atrial myocytes remains controversial (see above), likely due to the technically difficult evaluation of this Ca²⁺ release mechanism. Further studies are clearly warranted and might help to elucidate Ca²⁺ release properties in different AF aetiologies.

Low [Na⁺]_i

[Na⁺]_i is an important regulator of intracellular Ca²⁺ homeostasis and altered [Na⁺]_i underlies alterations in intracellular Ca²⁺ handling in heart disease (e.g. in heart failure) (Pogwizd & Bers, 2004). While this is well known, data regarding [Na⁺]_i homeostasis in atrial disease remain sparse. The first quantitative measurements of [Na⁺]_i in viable atrial myocytes were provided by our study as a part of the calcium signalling silencing mechanism. Resting [Na⁺]_i in normal rabbit atrial myocytes is low (6.86 ± 0.36 mmol l⁻¹) and was reduced in rapid rate‐induced calcium signalling silencing (4.02 ± 0.49 mmol l⁻¹) (Greiser et al. 2014). The low [Na⁺]_i appears to be a consequence of well‐known features of sustained rapid atrial activation: reduction of I _Ca,L and shortening of the atrial action potential duration (Greiser & Schotten, 2013). Reduced I _Ca,L combined with a short action potential duration leads to a substantial reduction in Ca²⁺ entry in these cells. Consequently, Na⁺ entry via NCX is also reduced when sarcolemmal Ca²⁺ is in pump–leak balance. Calculations of Ca²⁺ and Na⁺ membrane fluxes over one activation cycle show that the reduced Ca²⁺ entry entails a significant reduction in Na⁺ entry through the NCX. During 1 Hz stimulation 4 times as much Na⁺ enters the cell through the NCX than through Na⁺ channels (Despa et al. 2002). At rest Despa et al. reported a 100% higher Na⁺ influx through NCX than via Na⁺ channels in rabbit ventricular myocytes (Despa et al. 2002). The substantial reduction of Ca²⁺ entry through I _Ca,L and Ca²⁺ extrusion (via NCX) together with the coupled reduction in Na⁺ entry are likely the main mechanism underlying the low [Na]_i in calcium signalling silencing.

Thus, the unexpected and novel mechanism of low [Na⁺]_i‐mediated Ca²⁺ ‘unloading’ as a consequence of rapid rate underlines the importance of Na⁺ homeostasis in atrial disease. More specifically, increases in late Na⁺ current have been reported in atrial myocytes from AF patients and have been suggested to underlie Ca²⁺‐based arrhythmogenesis in AF (Sossalla et al. 2010). However, another study showed that the increases in the late Na⁺ current in myocytes from AF patients disappeared when recordings were performed at 37°C (Poulet et al. 2015). Given the importance of [Na⁺]_i homeostasis for Ca²⁺‐based arrhythmogenesis, future studies focusing on atrial [Na⁺]_i handling in AF are needed. Atrial myocytes from patients in AF are likely to show a degree of Na⁺ unloading due to calcium signalling silencing (as a consequence of the rapid rate).

Role of T‐tubules

In cardiac myocytes, T‐tubules (TTs) are sarcolemmal invaginations of the surface membrane that occur at the Z‐line and have transverse and longitudinal elements (Soeller & Cannell, 1999). Ventricular myocytes have abundant TTs with regular organization (Soeller & Cannell, 1999). This leads to a close proximity of LTCCs to RyR2s in these cells and facilitates a near simultaneous increase in [Ca²⁺]_i after membrane depolarization. In atrial myocytes TT density is lower and TTs are less well organized. In addition, there is evidence for differences in the LTCC microdomain. In a recent study LTCCs in rat ventricular myocytes were found to be located in the TTs, whereas in rat atrial myocytes ∼30% of LTCCs were located in caveolae. Interestingly, reduction of caveolae in atrial cells reduced the frequency of spontaneous calcium release events in these cells (Glukhov et al. 2015). This finding suggests that differences in LTCC–TT microdomains contribute to the differences between atrial and ventricular [Ca²⁺]_i signalling. Atrial TTs occur with higher abundance in sheep (Dibb et al. 2009; Lenaerts et al. 2009), dog (Wakili et al. 2010) and mice (Greiser et al. 2014). A mainly longitudinally oriented, substantial TT system has been described in rat atrial myocytes (Kirk et al. 2003). On the other hand, TTs are sparse in human, cat, rabbit and guinea‐pig atrial cells (Forbes et al. 1990; Berlin, 1995; Hüser et al. 1996; Richards et al. 2011; Greiser et al. 2014). Importantly, in these cells there is a significant delay between the subsarcolemmal and central‐cellular CaT, which is due to the lack of TTs (Hüser et al. 1996; Greiser et al. 2014). Atrial cells with abundant TTs, on the other hand, have a ‘ventricular’ pattern of intracellular Ca²⁺ release resulting in a homogeneous CaT. Indeed, the spatio‐temporal heterogeneity of the human atrial CaT and its relation to the lack of TT in these cells was already described in a seminal study by Hatem and coworkers (Hatem et al. 1997). Importantly, the amplitudes of the subsarcolemmal and central‐cellular CaT are similar during steady‐state field stimulation in rabbit and human atrial myocytes. Ca²⁺ signalling silencing induces a significant reduction of the central‐cellular CaT, without affecting the amplitude of the subsarcolemmal CaT (Greiser et al. 2014). This ‘silencing’ of the central‐cellular part of the CaT also underlies the reduced whole cell Ca²⁺ transient seen in atrial myocytes isolated from patients with AF (Greiser et al. 2014) showing that key aspects of Ca²⁺ signalling silencing prevail in AF. Reduced central CaTs have been previously described in AF models in sheep and dog atrial myocytes. The reduced central CaTs in these models were due to significant TT depletion, which reduced LTCC–RyR2 pairs (Dibb et al. 2009; Lenaerts et al. 2009; Wakili et al. 2010).

Thus, in species with functionally important atrial TTs (sheep, dog, mouse), TT depletion is an important mechanism of altered Ca²⁺ signalling in various models of atrial remodelling. In atrial cells with sparse TTs, calcium signalling silencing reduces central CaT amplitudes independently of changes in TT abundance. Further, isolated human left atrial myocytes from the posterior left atrium, a region close to the arrhythmogenic pulmonary vein sleeves, were found to have no functionally relevant TTs in either control or AF myocytes (Greiser et al. 2014). Should this subcellular anatomy be consistent throughout the left atrium, it would strongly suggest that TT depletion is not an important mechanism in human AF, but rather that rapid rate induced Ca²⁺ signalling silencing underlies the reduction of the central‐cellular CaT in human AF.

Relevance of Ca²⁺ silencing in the natural progression of AF

Calcium signalling silencing is the first ‘anti‐arrhythmic’ adaptive mechanism induced by a feature of AF that was thought to further the arrhythmia, namely rapid atrial activation. Sustained rapid rates that develop as a precursor to the arrhythmia as episodes of paroxysmal AF prevent the development of arrhythmogenic Ca²⁺ instability. It follows that the Ca²⁺ signalling ‘phenotype’ in a specific AF patient is the sum of all changes in intracellular Ca²⁺ signalling. Whether arrhythmogenic Ca²⁺ instability or anti‐arrhythmic silencing will prevail is likely to depend on the aetiology or ‘type’ of AF (Fig. 2). Heart failure, which predisposes to AF, leads to unstable arrhythmogenic Ca²⁺ instability in ventricular cells (Pogwizd & Bers, 2004). It is therefore likely that unstable Ca²⁺ signalling contributes to the initiation of AF in heart failure. Indeed, a canine heart failure model produced unstable atrial Ca²⁺ signalling resulting in increased DADs (Yeh et al. 2008). Similarly, the only transgenic mice that develop spontaneous sustained AF have a cardiomyopathy induced by high [Na⁺]_i (Wan et al. 2016), suggesting Ca²⁺‐based arrhythmogenesis as an underlying mechanism.

Figure 2. Schematic diagram for the proposed role of Ca²⁺ signalling silencing in Ca²⁺‐based arrhythmogenesis during the natural progression of different types of AF.

See text for details.

The effects of Ca²⁺ signalling silencing can persist in patients as shown in our data from human AF. The failure of the centripetal intracellular Ca²⁺ wave, which is a key feature of Ca²⁺ signalling silencing, is observed in field stimulated atrial cells from patients with persistent AF (Greiser et al. 2014). Similarly, reduced rather than enhanced spontaneous ectopic activity was reported in isolated atrial trabeculae from patients with persistent AF compared to trabeculae from patients in sinus rhythm (Christ et al. 2014).

It may well be that underlying cardiovascular disease (e.g. heart failure, ischaemic heart disease) produces a Ca²⁺‐based arrhythmogenic substrate and that Ca²⁺‐based arrhythmogenic mechanisms contribute to the initiation of AF in these patients. After the onset of AF, however, some degree of Ca²⁺ silencing will occur and Ca²⁺‐based arrhythmogenesis at this point might not further contribute to AF maintenance. Similarly, in patients without accompanying heart disease (‘lone AF’), Ca²⁺‐based arrhythmogenesis might not play a role in initiation and maintenance of AF (Fig. 2). Consistent with this concept is the fact that lone rapid atrial pacing or experimental AF as such has never been shown to produce an enhanced rate of atrial ectopic activity (Schotten et al. 2011).

In summary, future studies are needed that focus on the pathophysiological characteristics of different AF aetiologies. Arrhythmogenic unstable Ca²⁺ signalling, for example, is more likely to occur in AF developing in heart failure patients than in patients with lone atrial fibrillation, where Ca²⁺ signalling silencing might be dominant. Elucidating distinct changes in [Ca²⁺]_i and [Na⁺]_i homeostasis for different AF aetiologies will be crucial for developing ‘personalized’ therapeutic concepts for individual AF patients.

Additional information

Competing interests

None declared.

Funding

This work was supported by a Scientist Development Grant from the American Heart Association (14SDG20110054) to the author.

Biography

Maura Greiser trained in Internal Medicine/Cardiology in Aachen, where her interest in atrial fibrillation developed. She then did a PhD with Uli Schotten and Maurits Allessie in Maastricht. After that she moved to Baltimore to join Jon Lederer's group where her work continued to focus on atrial Ca²⁺ signalling in atrial fibrillation. She joined the faculty at the University of Maryland School of Medicine in the Department of Physiology.